Incentives can be created by either carrots (rewards) or sticks (penalties). By optimally combining them, a principal can reduce—at times, to zero—the distortions that incentives create for agents’ activity levels (entry or exit) in an environment with asymmetric information. A simple model is used to show that the scope for sticks widens when the dominant problem is imperfect information—the principal can imperfectly observe the agent’s actions—while it narrows when the dominant problem is incomplete information—the principal can imperfectly observe the agent’s type. This finding has implications for the optimal design of regulatory and contractual incentives.